Written by:
Executive Director, Haematology R&D Translational Medicine, AstraZeneca
Global Head of CLL Clinical Development and Strategy, AstraZeneca
For patients living with blood cancer, confirming that treatment has been successful is rarely straightforward. Even when there are no visible signs of disease, a crucial question remains: have all cancer cells been eliminated?
This matters because even a small number of cells remaining after treatment can multiply, resulting in disease relapses. These lingering cells are called minimal residual disease (MRD), and technological advances are enabling their detection with great sensitivity. Combined with the capability of artificial intelligence (AI), MRD testing has the potential to transform how blood cancers are diagnosed and how treatment response is measured, enabling more informed, data-driven decisions in the clinic.
Our vision is for MRD testing to become standard in haematology care, to inform diagnosis, prognosis and adaptive treatment therapies.
Why is MRD testing important in blood cancer treatment decisions?
MRD testing has emerged as a robust prognostic tool, providing deep insights into how well a patient is responding to a particular treatment, and with greater precision than standard assessments, such as imaging, allow. Evidence indicates that a greater reduction in MRD is associated with better long-term prognosis, making it a valuable tool for assessing response, identifying patients at high risk of relapse, and informing ongoing clinical decisions.
The field of MRD testing is rapidly evolving. The advent of highly sensitive methods like next-generation sequencing (NGS) has made it possible to identify tiny traces of cancer by measuring cells, DNA, protein, or other soluble materials present in blood and bone marrow. This can help identify patients who may appear MRD-negative by older methods, but who may benefit from additional treatment.
How does MRD testing benefit blood cancer patients and drug development?
There are three core applications that MRD testing offers. While these principles apply broadly across types of cancer, haematology is one of the areas where MRD is considered the most advanced. These include:
Prognosis
MRD testing can provide a window into prognosis at the end of therapy, identifying whether a patient is at risk of disease relapse, and may benefit from close monitoring or further treatment.
Informing treatment decisions
The depth and speed of treatment response is crucial as this can be an early indicator of outcome. MRD monitoring during treatment can provide clinicians with a clearer picture of how effectively a treatment is working, to inform adaptive treatment approaches. For example, patients with residual disease may benefit from additional or “consolidation” therapy, aiming to minimize their risk of relapse.
Supporting drug development
If MRD can serve as a surrogate endpoint for overall survival, clinical trials could read out more quickly, potentially bringing new treatment options to patients sooner. This potential is already being realised in multiple myeloma, where regulators have accepted MRD as an intermediate regulatory endpoint to support the accelerated approval of some classes of therapy. Because many current therapies in myeloma can achieve high response rates, MRD offers a sensitive way to identify those that drive deeper, more durable responses.
How is AstraZeneca using AI to advance MRD research?
AstraZeneca is actively advancing new approaches to realise the full potential of MRD through AI. It’s now possible to obtain large amounts of data from small biological samples, including information on DNA and methylation status and proteomics, and the challenge now is how to maximise this data. Through our collaborations, we have access to advanced foundational models capable of dissecting these rich datasets at a depth and scale that would not otherwise be possible.
Looking further ahead, we see AI playing three increasingly beneficial roles. The first is predictive, identifying which type of data is most informative for determining treatment response. The second is interpretive, using AI to determine directly whether a patient is MRD-positive or negative, with greater consistency and precision. The third, and most ambitious, is explanatory, to understand disease and identify mechanisms of treatment resistance that can help explain why a therapy may stop being effective.
How is MRD research helping advance the future of blood cancer care?
While MRD is not yet routinely accepted as a regulatory endpoint across all blood cancers, its use in clinical trials is growing rapidly. Together with partners across the ecosystem, we are advancing MRD by embedding it across the breadth of our haematology programs. This includes strategic partnerships with diagnostic leaders such as GRAIL, where we are exploring methylation-based biomarker approaches to assess MRD in the post-treatment setting, and our work with academic centres and diagnostic partners to run large-scale MRD testing across CLL, multiple myeloma and non-Hodgkin lymphoma.
We are also actively participating in consortia with industry, diagnostic and regulatory agencies, that are focused on establishing MRD as a routinely accepted regulatory endpoint. The technologies we are exploring today have the potential to become core aspects of haematology care of tomorrow, helping more patients with blood cancer receive the right treatment, at the right time, across every stage of their disease.
